Drone Delivery Packing Problem on a Neutral‐Atom Quantum Computer

ABSTRACT

We address the Drone Delivery Packing Problem (DDPP) via a hybrid quantum‐classical framework in which a neutral‐atom quantum processing unit (QPU) is used in analog mode as an independent‐set sampler, exploiting the Rydberg blockade mechanism. We reformulate DDPP as a graph‐partitioning problem based on the independent sets (ISs) of a scheduling graph encoding delivery incompatibilities. Each partition corresponds to deliveries assigned to a single drone, with the objective of minimizing the total number of partitions. ISs represent time‐feasible schedules, while battery constraints are enforced through a post‐processing routine. The success of the methodology depends on QPU sample‐set size and quality, with optimal schedules recoverable given a sufficiently high sample count. To analyze the effect of the sampling method, we benchmark the emulated quantum IS sampler (QIS) against a classical random‐greedy IS sampler (RG‐IS) within the same pipeline for DDPP, finding overall comparable solution quality. Experiments on Pasqal's Fresnel QPU reproduce the emulator performance on the tested hardware layouts up to 100 atoms, although a non‐negligible additive gap remains at the largest sizes. The results therefore establish the feasibility of integrating current neutral‐atom sampling into a hybrid workflow, rather than a systematic advantage over classical IS sampling.